论文部分内容阅读
目的:建立一种新的正确识别率更高的近红外光谱分析方法。方法:对传统的相关系数方法进行改进。以参考光谱特征性的峰和谷作为计算相关系数的依据,并使用琥乙红霉素片不同批次样品的光谱、在不同仪器上测得的光谱以及不同厂家样品的光谱等进行验证。结果:与传统的相关系数方法以及一致性检验模型相比,峰谷相关系数法的正确识别率更高。结论:峰谷相关系数法作为一种新方法,具有实用性,可以在近红外光谱药品快速检測技术中发挥一定的作用。
Objective: To establish a new near-infrared spectroscopy method with higher correct recognition rate. Methods: The traditional correlation coefficient method is improved. Based on the peaks and valleys characteristic of the reference spectra, the correlation coefficient was calculated, and the spectra of samples of different batches of erythromycin ethylsuccinate, the spectra measured on different instruments and the spectra of different manufacturers were validated. Results: Compared with the traditional correlation coefficient method and the consistency test model, the correlation coefficient between peaks and valleys is more correct. Conclusion: The peak-valley correlation coefficient method is a new method which has practicality and can play a certain role in the rapid detection of near-infrared spectroscopy.